Web analytics involves collecting and analysing a number of metrics to better understand the user experience offered by a website. For example, by using services such as Google Analytics it is possible to collect information on how long people spend on the website, how many people leave the website instantly, which websites people arrive from and where people are dropping out in key user journeys.
The analytics buzz
There is a growing buzz around how Web analytics can help optimise the online user experience. Web analytics can be useful in flagging potential issues and areas for further analysis. Indeed, Web analytics can be an integral part of the overall user centred design process. However, one of the biggest misunderstandings is that Web analytics alone can deliver accurate and actionable recommendations. Let’s look at an example.
A metric often used in Web analytics is bounce rate. This measures the percentage of people who come to your website and then leave almost instantly. At first glance, it would be reasonable to conclude that a higher bounce rate is worse than a lower bounce rate. So, imagine if your website has a high bounce rate. The first step should be to understand why it is happening.
Understanding the why
You may be able to dig a little deeper into the Web analytics you have. For example, there are typically two reasons why a high bounce rate can occur on a website. Firstly, your website may be attracting the wrong people. Your Web analytics will tell you which websites people were on before they got to yours. As a result, you can guess whether these websites are directing the right or wrong people to your website. If the referring website is in the same sector as yours, that’s probably fine. If you have an online cake shop and people were directed to your website from a Google search on cars, that’s an issue. Secondly, if the right people are going to your website, the high bounce rate may be explained by a poorly designed user experience. This could be caused by a variety of reasons. For example, having confusing navigation, content that does not engage or poorly designed call to actions. The issue here, however, is that no number and no graph from your Web analytics will be able to tell which one of these it is with any certainty. Essentially, an important missing piece with Web analytics is often the why.
Without knowing the why, there is a bigger danger of creating phantom issues. That is, issues that do not actually exist. Taking bounce rate again, because there is no way to know for certain why people are bouncing off the first page they visit, it is impossible to say whether a high bounce rate is even an issue. For example, people may have specific goals in mind when visiting your website. This could be reading the latest blog post or looking up contact information. To meet these goals, people may typically navigate directly to the piece of content they need, perhaps from a Web search or an email newsletter, and then leave. This should be seen as a successful interaction as the website successfully supported the person in their task. However, the high bounce rate resulting from this interaction would indicate an issue. Essentially, without knowing why the user visited, the Web analytics would suggest a phantom issue. Take another common metric from Web analytics, time spent on the website. At first glance, it would be reasonable to conclude the longer people spend on the website, the better. So imagine the Web analytics is reporting people are spending only a small amount of time on your website. As with the bounce rate example, the first step should be to understand why it is happening. Again, the trouble is there is nothing conclusive you can draw upon in your Web analytics to understand with certainty why people are only spending a short amount of time on your website. For example, people may only spend a few minutes on a website and yet complete their goals and have a satisfying experience. Although this is a successful interaction with a user, looking solely at the time it would appear to be an unsuccessful interaction. Additionally, it is wrong to conclude that a metric indicating someone has spent a long time on a page or website is always positive. For example, that person could be spending a long time on the website because they are either engaged or confused by content. Essentially, without knowing if it is an issue or not, it is impossible to create accurate and actionable recommendations for change.
Let’s put this in context. Imagine you want to understand the experience people are having in a restaurant you own. Based upon a percentage of how many people left your restaurant immediately and how long they stayed there, would you be able to accurately understand the experience your diners are having? Probably not. If people came in and left your restaurant immediately (or bounce rate on your website), you would not know why. It could be because there were no tables free, they did not like the menu or had to take a phone call outside. Additionally, there would be no way to understand for certain why people were spending a short time eating their meal (or the time spent on website). Perhaps they were late for a movie or perhaps they were just really hungry. You would not know which.
A suitable place for analytics
Reflecting on the examples with bounce rate and time spent on the website. Although Web analytics alone will not tell you why these issues are occurring and there is a risk of creating phantom issues, these metrics can be useful in flagging potential issues and areas for further analysis. Essentially, Web analytics can be an integral part of the overall user centred design process and strategy for improving the user experience, however, not the only part.
A clearer picture
To ensure the insight provided by Web analytics can be translated into accurate and actionable improvements to improve the user experience, it is essential to use other supporting research techniques. Usability testing, accessibility audits, expert usability reviews and information architecture research can answer the ‘why’. For example:
- Web analytics can tell you if your key calls to action are not effective, but not why. Usability testing, however, will provide a clear picture as to why;
- Web analytics can tell you if people are struggling with a form, but not why. An expert usability review on the form, however, will provide a clear picture as to why;
- Web analytics can tell you where people are dropping out in a journey, but not why. An accessibility audit, however, could identify key accessibility barriers.
Indeed, Web analytics can be used to increase the effectiveness of these techniques. For example, identifying potential issues with Web analytics first will enable you to focus your efforts more effectively when doing an expert usability review.
Although Web analytics can be an integral part of the overall user centred design process and strategy for improving the user experience, it is not the only part. Ensure you follow up any potential issues raised with other user research, such as usability or accessibility testing and information architecture (IA) research.
Can we help?
UX Research is at the very heart of what we do and a key component in any work we deliver. Getting the research right at the start of your project will ensure that your project will run smoothly and meet the needs of your customers.
If you would like to speak to us about how we can help you with your UX research then please do not hesitate to give us a call 44 (0) 117 929 7333 or submit this short form. in the meantime please take a look at our UX research services.